Residential Plumbing Retrofits

The savings estimates presented below are based on a series of rigorous field studies that examined the change in metered water consumption of more than 27,000 households and customers in the Cities of Irvine, Los Angeles, San Diego, and Santa Monica. Because the exact number and type of devices contained in a retrofit kit can and has varied significantly, device-level estimates assist the comparison across studies.

Device/Activity Description

Low flow (LF) showerheads are designed to provide water at lower rates of water flow. Flow is typically measured in gallons per minute and low flow showerheads are rated at 2.5 gallons per minute (gpm) or less (at pressure levels up to 80 psi). California state law currently requires that all showerheads sold in the state meet the 2.5 gpm standard. Toilet displacement devices come in a variety of designs that displace some water volume in the toilet tank. Since less water is needed to refill the tank, less water is used per flush. Toilet leak detection is typically performed with dye tablets. Faucet aerators reduce flow from faucets.

Applicable BMPs

BMP 1 – Residential Water Surveys. Residential surveys may involve plumbing retrofits.

BMP 2 – Residential Plumbing Retrofit.

Available Water Savings Estimates

Showerheads
The water savings estimates below represent a statistical estimate of the mean change in water use observed over a large number of residential households. We present a subset of estimates from these field studies that: (1) are based on a large sample size, (2) represent a multiple year period, and (3) have statistically controlled for non-plumbing related factors and ongoing conservation. It is desirable to have a large sample size so as to increase the precision of the estimate. A multiple year period is needed to examine patterns over time. Careful control for biasing effects is required to ensure the estimates represent net water savings, not gross water savings—that is, savings from conservation programs, not from other factors such as household characteristics. Table 1 provides a summary of these estimates.


Table 1 - Statistical Estimates of Low Flow Showerhead Savings

Estimates Margin Time Period Sample Size Source
5.5 gpd/LFSH Single Family +/- 1.5 gpd 1990-92 ~2,900 SF Dwellings (3)
5.8 gpd/LFSH Single Family +/- 2.6 gpd 1990-93 ~3,000ngs SF Dwelli (4)
5.2 gpd/LFSH Multi-Family +/- 1.1 gpd 1990-92 ~2,300 MF Complexes (9.5 Units/Complex) (3)


The probability of a showerhead actually being replaced can vary widely. The probability of replacement depends in part on the method of distribution (e.g., “hang and pray”). Field studies of retrofit kit distributions in Irvine (Chesnutt et al. 1992) and Los Angeles (Chesnutt et al. 1991) have found initial installation probabilities that range from 49 percent to 59 percent. Not all showerheads that are replaced are retained. Since both estimates reflect self-reports, they may overstate the true installation probability. The same two field studies found that 7-9 percent of installed LF showerheads were later removed. Direct install programs allow a direct count of the number of installed showerheads; only the probability of removal then needs to be estimated.

MWDSC and MWDOC (2002) report the results of an extensive study of the saturation of conservation devices in Orange County. The study found that countywide surveys of low flow showerhead saturation provide good estimates of the saturation of the individual agencies within the county. Saturation was found to be near 75 percent.

Aquacraft (1999) reports that low flow shower homes, “used an average of 29.9 gpd and 11.1 gpcd for showering, while the non-LF shower homes used an average of 34.4 gpd and 13.3 gpcd. The study reports a statistically significant difference of 2.2 gpcd on average.

Other Devices
Table 2 shows water savings estimates for other plumbing retrofit devices from a field study in Los Angeles (Chesnutt et al. 1995b). Even with the large sample size of this study, these estimates of the expected change in metered household water consumption are less precise than the showerhead estimates. In the two field studies of plumbing retrofit programs mentioned above (Chesnutt et al. 1991 and 1992), toilet dams exhibited somewhat higher self-reported installation rates and higher removal rates. Estimates of the installation rate for faucet aerators also come from self-reported data and, as such, should also be considered speculative. The field study in Irvine Ranch found that 13 percent of respondents reported the use of leak detection tablets. Estimates of the rate of toilet leakage derive from Bamezai and Chesnutt (1994), Chesnutt et al. 1995b, Chesnutt et al. 1991, and Steirer and Broder 1994.


Table 2 - Statistical Estimates of Savings from Other Retrofit Devices

Retrofit Device Savings (gpd/device) Error Margin
Toilet Dams 4.2 gpd +/- 2.6 gpd
Faucet Aerators 1.5 gpd +/- 2.6 gpd
Leak Detection Tablets 8 gpd +/- 2.6 gpd


CUWCC (2002) reports results from an MWD study: “installation of the wrong flapper in a water efficiency fixtures could result in water consumption amounts as high as 4.4 gallons for non­adjustable flappers and 3.4 gallons for adjustable flappers.”

Koeller (2004) reports that only 11 percent of the studied toilet fixtures had leaking flapper valves. The study also found, “approximately 52 percent of all aging toilet fixtures inspected are flushing at a rate of 1.7 gpf or higher.” Approximately 12 percent were flushing below 1.4 gpf.

Persistence Showerhead savings estimates have been measured in recent programs. Since these field studies examined water use over a multi-year period, the estimates reflect the multi-year period average and they embed any retention and decay effects. There is some evidence that future programs may yield less water savings due to the increasing saturation of LF showerheads in most service areas. State plumbing code requiring sale of LF showerheads tends to increase the saturation of low flow showerheads over time. Direct evidence of background saturation rates can be derived from data collected during home water surveys. Table 3 shows flow rates of existing showerheads as measured in recent residential surveys in Los Angeles and San Diego.


Table 3 - Flow Rate of Existing Showerheads

Home Survey Location Flow Rate of Existing Showerheads Time Period Sample Size Source
Los Angeles 3 gpm Summer 1993 5,502 SF Residences (10)
San Diego 3.08 gpm FY 1994-95 3,666 SF Residences and 489 MF (11


Limitations
Since conserving showerheads are required in plumbing code, background saturation rates are likely to be higher now than during the study periods referred to above.

Confidence in Estimates
Medium to High.

Considerable empirical research has been conducted regarding the savings of low flow showerheads. Important areas for future research include background saturation rates and persistence of savings over time.

Program and Device/Activity Cost Estimates

Program Costs

Participant program costs may include:

  • Cost of retrofit kit if not fully subsidized
  • Installation cost if not fullysubsidized Supplier program costs may include:
  • Staff time to contact building departments, developers, and plumbing supply outlets
  • Retrofit kits: showerheads, toilet displacement devices, and installation costs
  • Administration
  • Contractors
  • Marketing


The following are professional judgments of costs by conservation program coordinators and managers, as reported in A&N Technical Services (1995):

  • Low flow showerheads, kit: $2
  • Low flow showerheads, direct install: $10-15


Limitations
Cost estimates vary with the scale of the program.


==Confidence in Estimates==
Medium.

Water Savings Calculation Formula(s)

Calculations Water Savings = Device_Savings * Number_of_Devices * Probability_of_Installation * Lifespan

Factors to Consider in Applying the Formula
Per device water savings from field studies embed behavioral responses (longer showering times) and mechanical/engineering estimates do not. Water savings decay can be very site specific. Water supplies with high mineral content can degrade showerheads relatively quickly.

This affects the background saturation rate, degradation of new showerheads, and ongoing device replacement rates. The probability of installation/retention is both site-specific and uncertain.

Example Calculations
Table 4 summarizes savings rates, life spans and decay rates for low flow showerheads and other retrofit devices. Method 1 is a method to account for savings decay by accounting for the savings over a number of years representing the device life span. Method 2 is an alternative method, whereby the savings are reduced by the indicated percent over the period of analysis (percent year over year, exponential) or until savings approach zero.


Table 4 - Retrofit Device Savings

Method 1 Method 2
Device Initial Savings (gpd per device) Device Life Span Device Decay Rate per Year
Low Flow Showerheads 5.5 gpd 3-7 years 20-30 percent
Toilet Displacement Devices 4 gpd 2-5 years 40-60 percent
Faucet Aerators 1.5 gpd 1-3 years 40-60 percent
Toilet Leak Detection .64 gpd (8 gpd per repaired leaking toilet; 8 percent of toilets leak) 7-10 years 1-2 percent
Other Household Leak Check .5 gpd (12.4 gpd per household repair; 4 percent of households with leaks) 7-10 years 1-2 percent
Turf Audit 12.2 4 years 40-60 percent
Turf Audit with Timer 25.9 gpd (12.2 gpd for turf audit plus 13.7 if timer) 4 years 40-60 percent
Source Field Studies Judgment Judgment


Questions to Ask

  • Are there other agencies that you can partner with to make your program more cost effective?
  • Does your agency have access to grant or other partnership type funding?
  • Are devices to be provided on “hang and pray” or “directly installed” basis?
  • Will the selected method be accomplished with agency’s own personnel or using a contractor?
  • Does your agency allow your agency personnel or contractor personnel to enter the customer’s home?
  • What marketing technique will be used to accomplish the selected method?
  • What devices and actions are included?
  • Will your personnel or the contractor’s personnel install the devices? If not, how will installations be verified?
  • Do you have estimated or comparative cost for device components and method selected to implement the program?
  • Are you going to design and maintain a database covering program results?
  • What is the age of the housing stock?
  • Can you influence how the cost of this program is accounted for? If capitalized, the cost impact will be spread over “x” number of years and reduce the rate impact. If expensed, will the cost of your program have to be recovered in one year?


Sources

Aquacraft, Inc., “Residential End Uses of Water,” prepared for American Water Works Research Foundation, 1999.

Bamezai, A. and T.W. Chesnutt, Residential Water Audit Program: Evaluation of Program Outcomes and Water Savings, A report for the Metropolitan Water District of Southern California, December 1994.

Brown and Caldwell, Residential Water Conservation Projects, Summary Report, U.S. Department of Housing and Urban Development, June 1984.

Chesnutt, T.W., C.N. McSpadden, and A. Bamezai, Ultra Low Flush Toilet Programs: Evaluation of Program Outcomes and Water Savings, A report for the Metropolitan Water District of Southern California, July 1995.

Chesnutt, T.W., C.N. McSpadden, and D.M. Pekelney, What is the Reliable Yield from Residential Home Water Survey Programs?, Presented at the AWWA Conference in Anaheim CA, June 1995b.

Chesnutt, T.W., C. N. McSpadden, S. A. Adnan, and A. Bamezai, A Model-Based Evaluation of Irvine Ranch Water District Residential Retrofit and Survey Water Conservation Projects, A report for the Metropolitan Water District of Southern California, August 1992.

Chesnutt, T.W., Bamezai, A., and C.N. McSpadden, Continuous-Time Error Components Models of Residential Water Demand, A report for the Metropolitan Water District of Southern California, June 1992.

Chesnutt, T.W. and C.N. McSpadden, A Model-Based Evaluation of the Westchester Water Conservation Programs, A report for the Metropolitan Water District of Southern California, January 1991.

Chesnutt, T.W. and C.N. McSpadden, The Evaluation of Water Conservation Programs: What is Wrong with the Industry Standard Approach?, A report for the Metropolitan Water District of Southern California, January 1991.

Chesnutt, T.W. and C.N. McSpadden, Improving the Evaluation of Water Conservation Programs, A report for the Metropolitan Water District of Southern California, January 1991.

Hahm, W. and T.W. Chesnutt, Data Used in the Evaluation of the Los Angeles Department of Water and Power Home Water Survey, A report for the Los Angeles Department of Water and Power, September 1994.

CUWCC (2002), “Toilet Flappers: A Weak Link In Water Conservation,” paper prepared by J. Koeller for the California Urban Water Conservation Council, March.

Metropolitan Water District of Southern California (MWDSC) and the Municipal Water District of Orange County (MWDOC), “ Orange County Saturation Study,” July 2002.

Steirer, M. A. and M. I. Broder, Residential Water Survey Program Final Report for Fiscal Year 1994-95, Prepared by the City of San Diego Water Utilities Department Water Conservation Program, November, 1995.

Pekelney, D.M., and T.W. Chesnutt, “Reference Document: Program Design Tool and Savings Estimates,” A&N Technical Services (1995), for MWDSC.



Has general subjectBenefits and costs +
Has introductionThe savings estimates presented below are
The savings estimates presented below are based on a series of rigorous field studies that examined the change in metered water consumption of more than 27,000 households and customers in the Cities of Irvine, Los Angeles, San Diego, and Santa Monica. Because the exact number and type of devices contained in a retrofit kit can and has varied significantly, device-level estimates assist the comparison across studies.
ates assist the comparison across studies. +